I'm trying to solve this problem
Create a sampling distribution for our 30-observation-mean estimator of the mean of C. To do so, randomly draw 1000 sets of 30 observations from C, using the randomly drawn indices in D (the first row are first 30 randomly-drawn indexes, the second row are the second 30 randomly-drawn indexes, etc). For each random draw, compute the mean. Then plot the histogram of the distribution. Compare the distribution to np.mean(C).
Where C is
array([23, 23, 23, ..., 68, 34, 42])
and size of C is 100030 and D column is (with size 30000)
array([[23989, 10991, 81533, ..., 75050, 13817, 47678],
[54864, 54830, 89396, ..., 22709, 14556, 62298],
[ 2936, 28729, 4404, ..., 21431, 81187, 49178],
...,
[30737, 12974, 41031, ..., 43003, 61132, 33385],
[64713, 53207, 49529, ..., 72596, 76406, 15207],
[29503, 71648, 27210, ..., 31298, 47102, 13024]])
I'm trying to understand the problem here and how to solve it. What I have done so far is initializing a list with zeros and trying to get the mean based on the indices in D. But I'm not sure if this is what is actually asked for? Any help?
samp = np.zeros( (1000, 1))
for i in np.arrange(0, 1000):
samp(i) = np.mean(C( D(i,)))
also, this is taking random samples from C but not sure how to add D indices to it?
means_size_30 = []
for x in range(1000):
mean = np.random.choice(C, size = 30).mean()
means_size_30.append(mean)
means_size_30 = np.array(means_size_30)
plt.hist(means_size_30);
You can directly access the values of C by using the indexes provided in D. If you use the 2-dimensional array D to access values of the 1-dimensional array C, the resulting array will have the same shape as D: 2-dimensional. It will have 1000 rows with each row having 30 samples from C.
In the next step you just have to calculate the mean over each row (set axis=1):
means_size_30 = C[D].mean(axis=1)
plt.hist(means_size_30)
plt.axvline(np.mean(C))